exec.c 43.2 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
/*
 *  linux/fs/exec.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 */

/*
 * #!-checking implemented by tytso.
 */
/*
 * Demand-loading implemented 01.12.91 - no need to read anything but
 * the header into memory. The inode of the executable is put into
 * "current->executable", and page faults do the actual loading. Clean.
 *
 * Once more I can proudly say that linux stood up to being changed: it
 * was less than 2 hours work to get demand-loading completely implemented.
 *
 * Demand loading changed July 1993 by Eric Youngdale.   Use mmap instead,
 * current->executable is only used by the procfs.  This allows a dispatch
 * table to check for several different types  of binary formats.  We keep
 * trying until we recognize the file or we run out of supported binary
 * formats. 
 */

#include <linux/slab.h>
#include <linux/file.h>
A
Al Viro 已提交
27
#include <linux/fdtable.h>
H
Hugh Dickins 已提交
28
#include <linux/mm.h>
L
Linus Torvalds 已提交
29 30 31
#include <linux/stat.h>
#include <linux/fcntl.h>
#include <linux/smp_lock.h>
H
Hugh Dickins 已提交
32
#include <linux/swap.h>
33
#include <linux/string.h>
L
Linus Torvalds 已提交
34
#include <linux/init.h>
35
#include <linux/pagemap.h>
36
#include <linux/perf_event.h>
L
Linus Torvalds 已提交
37 38 39 40 41 42
#include <linux/highmem.h>
#include <linux/spinlock.h>
#include <linux/key.h>
#include <linux/personality.h>
#include <linux/binfmts.h>
#include <linux/utsname.h>
43
#include <linux/pid_namespace.h>
L
Linus Torvalds 已提交
44 45 46 47 48
#include <linux/module.h>
#include <linux/namei.h>
#include <linux/proc_fs.h>
#include <linux/mount.h>
#include <linux/security.h>
M
Mimi Zohar 已提交
49
#include <linux/ima.h>
L
Linus Torvalds 已提交
50
#include <linux/syscalls.h>
51
#include <linux/tsacct_kern.h>
M
Matt Helsley 已提交
52
#include <linux/cn_proc.h>
A
Al Viro 已提交
53
#include <linux/audit.h>
R
Roland McGrath 已提交
54
#include <linux/tracehook.h>
J
Johannes Berg 已提交
55
#include <linux/kmod.h>
56
#include <linux/fsnotify.h>
57
#include <linux/fs_struct.h>
L
Linus Torvalds 已提交
58 59 60

#include <asm/uaccess.h>
#include <asm/mmu_context.h>
61
#include <asm/tlb.h>
62
#include "internal.h"
L
Linus Torvalds 已提交
63 64

int core_uses_pid;
65
char core_pattern[CORENAME_MAX_SIZE] = "core";
A
Alan Cox 已提交
66 67
int suid_dumpable = 0;

L
Linus Torvalds 已提交
68 69
/* The maximal length of core_pattern is also specified in sysctl.c */

A
Alexey Dobriyan 已提交
70
static LIST_HEAD(formats);
L
Linus Torvalds 已提交
71 72
static DEFINE_RWLOCK(binfmt_lock);

I
Ivan Kokshaysky 已提交
73
int __register_binfmt(struct linux_binfmt * fmt, int insert)
L
Linus Torvalds 已提交
74 75 76 77
{
	if (!fmt)
		return -EINVAL;
	write_lock(&binfmt_lock);
I
Ivan Kokshaysky 已提交
78 79
	insert ? list_add(&fmt->lh, &formats) :
		 list_add_tail(&fmt->lh, &formats);
L
Linus Torvalds 已提交
80 81 82 83
	write_unlock(&binfmt_lock);
	return 0;	
}

I
Ivan Kokshaysky 已提交
84
EXPORT_SYMBOL(__register_binfmt);
L
Linus Torvalds 已提交
85

86
void unregister_binfmt(struct linux_binfmt * fmt)
L
Linus Torvalds 已提交
87 88
{
	write_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
89
	list_del(&fmt->lh);
L
Linus Torvalds 已提交
90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
	write_unlock(&binfmt_lock);
}

EXPORT_SYMBOL(unregister_binfmt);

static inline void put_binfmt(struct linux_binfmt * fmt)
{
	module_put(fmt->module);
}

/*
 * Note that a shared library must be both readable and executable due to
 * security reasons.
 *
 * Also note that we take the address to load from from the file itself.
 */
106
SYSCALL_DEFINE1(uselib, const char __user *, library)
L
Linus Torvalds 已提交
107
{
108 109 110 111
	struct file *file;
	char *tmp = getname(library);
	int error = PTR_ERR(tmp);

112 113 114 115 116 117 118 119 120
	if (IS_ERR(tmp))
		goto out;

	file = do_filp_open(AT_FDCWD, tmp,
				O_LARGEFILE | O_RDONLY | FMODE_EXEC, 0,
				MAY_READ | MAY_EXEC | MAY_OPEN);
	putname(tmp);
	error = PTR_ERR(file);
	if (IS_ERR(file))
L
Linus Torvalds 已提交
121 122 123
		goto out;

	error = -EINVAL;
124
	if (!S_ISREG(file->f_path.dentry->d_inode->i_mode))
L
Linus Torvalds 已提交
125 126
		goto exit;

127
	error = -EACCES;
128
	if (file->f_path.mnt->mnt_flags & MNT_NOEXEC)
L
Linus Torvalds 已提交
129 130
		goto exit;

131 132
	fsnotify_open(file->f_path.dentry);

L
Linus Torvalds 已提交
133 134 135 136 137
	error = -ENOEXEC;
	if(file->f_op) {
		struct linux_binfmt * fmt;

		read_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
138
		list_for_each_entry(fmt, &formats, lh) {
L
Linus Torvalds 已提交
139 140 141 142 143 144 145 146 147 148 149 150 151
			if (!fmt->load_shlib)
				continue;
			if (!try_module_get(fmt->module))
				continue;
			read_unlock(&binfmt_lock);
			error = fmt->load_shlib(file);
			read_lock(&binfmt_lock);
			put_binfmt(fmt);
			if (error != -ENOEXEC)
				break;
		}
		read_unlock(&binfmt_lock);
	}
152
exit:
L
Linus Torvalds 已提交
153 154 155 156 157
	fput(file);
out:
  	return error;
}

158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179
#ifdef CONFIG_MMU

static struct page *get_arg_page(struct linux_binprm *bprm, unsigned long pos,
		int write)
{
	struct page *page;
	int ret;

#ifdef CONFIG_STACK_GROWSUP
	if (write) {
		ret = expand_stack_downwards(bprm->vma, pos);
		if (ret < 0)
			return NULL;
	}
#endif
	ret = get_user_pages(current, bprm->mm, pos,
			1, write, 1, &page, NULL);
	if (ret <= 0)
		return NULL;

	if (write) {
		unsigned long size = bprm->vma->vm_end - bprm->vma->vm_start;
180 181 182 183 184 185 186 187
		struct rlimit *rlim;

		/*
		 * We've historically supported up to 32 pages (ARG_MAX)
		 * of argument strings even with small stacks
		 */
		if (size <= ARG_MAX)
			return page;
188 189 190 191 192 193 194 195

		/*
		 * Limit to 1/4-th the stack size for the argv+env strings.
		 * This ensures that:
		 *  - the remaining binfmt code will not run out of stack space,
		 *  - the program will have a reasonable amount of stack left
		 *    to work from.
		 */
196
		rlim = current->signal->rlim;
197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226
		if (size > rlim[RLIMIT_STACK].rlim_cur / 4) {
			put_page(page);
			return NULL;
		}
	}

	return page;
}

static void put_arg_page(struct page *page)
{
	put_page(page);
}

static void free_arg_page(struct linux_binprm *bprm, int i)
{
}

static void free_arg_pages(struct linux_binprm *bprm)
{
}

static void flush_arg_page(struct linux_binprm *bprm, unsigned long pos,
		struct page *page)
{
	flush_cache_page(bprm->vma, pos, page_to_pfn(page));
}

static int __bprm_mm_init(struct linux_binprm *bprm)
{
227
	int err;
228 229 230 231 232
	struct vm_area_struct *vma = NULL;
	struct mm_struct *mm = bprm->mm;

	bprm->vma = vma = kmem_cache_zalloc(vm_area_cachep, GFP_KERNEL);
	if (!vma)
233
		return -ENOMEM;
234 235 236 237 238 239 240 241 242 243 244 245 246

	down_write(&mm->mmap_sem);
	vma->vm_mm = mm;

	/*
	 * Place the stack at the largest stack address the architecture
	 * supports. Later, we'll move this to an appropriate place. We don't
	 * use STACK_TOP because that can depend on attributes which aren't
	 * configured yet.
	 */
	vma->vm_end = STACK_TOP_MAX;
	vma->vm_start = vma->vm_end - PAGE_SIZE;
	vma->vm_flags = VM_STACK_FLAGS;
247
	vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
248
	err = insert_vm_struct(mm, vma);
249
	if (err)
250 251 252 253 254 255 256
		goto err;

	mm->stack_vm = mm->total_vm = 1;
	up_write(&mm->mmap_sem);
	bprm->p = vma->vm_end - sizeof(void *);
	return 0;
err:
257 258 259
	up_write(&mm->mmap_sem);
	bprm->vma = NULL;
	kmem_cache_free(vm_area_cachep, vma);
260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358
	return err;
}

static bool valid_arg_len(struct linux_binprm *bprm, long len)
{
	return len <= MAX_ARG_STRLEN;
}

#else

static struct page *get_arg_page(struct linux_binprm *bprm, unsigned long pos,
		int write)
{
	struct page *page;

	page = bprm->page[pos / PAGE_SIZE];
	if (!page && write) {
		page = alloc_page(GFP_HIGHUSER|__GFP_ZERO);
		if (!page)
			return NULL;
		bprm->page[pos / PAGE_SIZE] = page;
	}

	return page;
}

static void put_arg_page(struct page *page)
{
}

static void free_arg_page(struct linux_binprm *bprm, int i)
{
	if (bprm->page[i]) {
		__free_page(bprm->page[i]);
		bprm->page[i] = NULL;
	}
}

static void free_arg_pages(struct linux_binprm *bprm)
{
	int i;

	for (i = 0; i < MAX_ARG_PAGES; i++)
		free_arg_page(bprm, i);
}

static void flush_arg_page(struct linux_binprm *bprm, unsigned long pos,
		struct page *page)
{
}

static int __bprm_mm_init(struct linux_binprm *bprm)
{
	bprm->p = PAGE_SIZE * MAX_ARG_PAGES - sizeof(void *);
	return 0;
}

static bool valid_arg_len(struct linux_binprm *bprm, long len)
{
	return len <= bprm->p;
}

#endif /* CONFIG_MMU */

/*
 * Create a new mm_struct and populate it with a temporary stack
 * vm_area_struct.  We don't have enough context at this point to set the stack
 * flags, permissions, and offset, so we use temporary values.  We'll update
 * them later in setup_arg_pages().
 */
int bprm_mm_init(struct linux_binprm *bprm)
{
	int err;
	struct mm_struct *mm = NULL;

	bprm->mm = mm = mm_alloc();
	err = -ENOMEM;
	if (!mm)
		goto err;

	err = init_new_context(current, mm);
	if (err)
		goto err;

	err = __bprm_mm_init(bprm);
	if (err)
		goto err;

	return 0;

err:
	if (mm) {
		bprm->mm = NULL;
		mmdrop(mm);
	}

	return err;
}

L
Linus Torvalds 已提交
359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374
/*
 * count() counts the number of strings in array ARGV.
 */
static int count(char __user * __user * argv, int max)
{
	int i = 0;

	if (argv != NULL) {
		for (;;) {
			char __user * p;

			if (get_user(p, argv))
				return -EFAULT;
			if (!p)
				break;
			argv++;
375
			if (i++ >= max)
L
Linus Torvalds 已提交
376 377 378 379 380 381 382 383
				return -E2BIG;
			cond_resched();
		}
	}
	return i;
}

/*
384 385 386
 * 'copy_strings()' copies argument/environment strings from the old
 * processes's memory to the new process's stack.  The call to get_user_pages()
 * ensures the destination page is created and not swapped out.
L
Linus Torvalds 已提交
387
 */
A
Adrian Bunk 已提交
388 389
static int copy_strings(int argc, char __user * __user * argv,
			struct linux_binprm *bprm)
L
Linus Torvalds 已提交
390 391 392
{
	struct page *kmapped_page = NULL;
	char *kaddr = NULL;
393
	unsigned long kpos = 0;
L
Linus Torvalds 已提交
394 395 396 397 398 399 400 401
	int ret;

	while (argc-- > 0) {
		char __user *str;
		int len;
		unsigned long pos;

		if (get_user(str, argv+argc) ||
402
				!(len = strnlen_user(str, MAX_ARG_STRLEN))) {
L
Linus Torvalds 已提交
403 404 405 406
			ret = -EFAULT;
			goto out;
		}

407
		if (!valid_arg_len(bprm, len)) {
L
Linus Torvalds 已提交
408 409 410 411
			ret = -E2BIG;
			goto out;
		}

412
		/* We're going to work our way backwords. */
L
Linus Torvalds 已提交
413
		pos = bprm->p;
414 415
		str += len;
		bprm->p -= len;
L
Linus Torvalds 已提交
416 417 418 419 420

		while (len > 0) {
			int offset, bytes_to_copy;

			offset = pos % PAGE_SIZE;
421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436
			if (offset == 0)
				offset = PAGE_SIZE;

			bytes_to_copy = offset;
			if (bytes_to_copy > len)
				bytes_to_copy = len;

			offset -= bytes_to_copy;
			pos -= bytes_to_copy;
			str -= bytes_to_copy;
			len -= bytes_to_copy;

			if (!kmapped_page || kpos != (pos & PAGE_MASK)) {
				struct page *page;

				page = get_arg_page(bprm, pos, 1);
L
Linus Torvalds 已提交
437
				if (!page) {
438
					ret = -E2BIG;
L
Linus Torvalds 已提交
439 440 441
					goto out;
				}

442 443
				if (kmapped_page) {
					flush_kernel_dcache_page(kmapped_page);
L
Linus Torvalds 已提交
444
					kunmap(kmapped_page);
445 446
					put_arg_page(kmapped_page);
				}
L
Linus Torvalds 已提交
447 448
				kmapped_page = page;
				kaddr = kmap(kmapped_page);
449 450
				kpos = pos & PAGE_MASK;
				flush_arg_page(bprm, kpos, kmapped_page);
L
Linus Torvalds 已提交
451
			}
452
			if (copy_from_user(kaddr+offset, str, bytes_to_copy)) {
L
Linus Torvalds 已提交
453 454 455 456 457 458 459
				ret = -EFAULT;
				goto out;
			}
		}
	}
	ret = 0;
out:
460 461
	if (kmapped_page) {
		flush_kernel_dcache_page(kmapped_page);
L
Linus Torvalds 已提交
462
		kunmap(kmapped_page);
463 464
		put_arg_page(kmapped_page);
	}
L
Linus Torvalds 已提交
465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482
	return ret;
}

/*
 * Like copy_strings, but get argv and its values from kernel memory.
 */
int copy_strings_kernel(int argc,char ** argv, struct linux_binprm *bprm)
{
	int r;
	mm_segment_t oldfs = get_fs();
	set_fs(KERNEL_DS);
	r = copy_strings(argc, (char __user * __user *)argv, bprm);
	set_fs(oldfs);
	return r;
}
EXPORT_SYMBOL(copy_strings_kernel);

#ifdef CONFIG_MMU
483

L
Linus Torvalds 已提交
484
/*
485 486 487
 * During bprm_mm_init(), we create a temporary stack at STACK_TOP_MAX.  Once
 * the binfmt code determines where the new stack should reside, we shift it to
 * its final location.  The process proceeds as follows:
L
Linus Torvalds 已提交
488
 *
489 490 491 492 493 494
 * 1) Use shift to calculate the new vma endpoints.
 * 2) Extend vma to cover both the old and new ranges.  This ensures the
 *    arguments passed to subsequent functions are consistent.
 * 3) Move vma's page tables to the new range.
 * 4) Free up any cleared pgd range.
 * 5) Shrink the vma to cover only the new range.
L
Linus Torvalds 已提交
495
 */
496
static int shift_arg_pages(struct vm_area_struct *vma, unsigned long shift)
L
Linus Torvalds 已提交
497 498
{
	struct mm_struct *mm = vma->vm_mm;
499 500 501 502 503 504
	unsigned long old_start = vma->vm_start;
	unsigned long old_end = vma->vm_end;
	unsigned long length = old_end - old_start;
	unsigned long new_start = old_start - shift;
	unsigned long new_end = old_end - shift;
	struct mmu_gather *tlb;
L
Linus Torvalds 已提交
505

506
	BUG_ON(new_start > new_end);
L
Linus Torvalds 已提交
507

508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533
	/*
	 * ensure there are no vmas between where we want to go
	 * and where we are
	 */
	if (vma != find_vma(mm, new_start))
		return -EFAULT;

	/*
	 * cover the whole range: [new_start, old_end)
	 */
	vma_adjust(vma, new_start, old_end, vma->vm_pgoff, NULL);

	/*
	 * move the page tables downwards, on failure we rely on
	 * process cleanup to remove whatever mess we made.
	 */
	if (length != move_page_tables(vma, old_start,
				       vma, new_start, length))
		return -ENOMEM;

	lru_add_drain();
	tlb = tlb_gather_mmu(mm, 0);
	if (new_end > old_start) {
		/*
		 * when the old and new regions overlap clear from new_end.
		 */
534
		free_pgd_range(tlb, new_end, old_end, new_end,
535 536 537 538 539 540 541 542
			vma->vm_next ? vma->vm_next->vm_start : 0);
	} else {
		/*
		 * otherwise, clean from old_start; this is done to not touch
		 * the address space in [new_end, old_start) some architectures
		 * have constraints on va-space that make this illegal (IA64) -
		 * for the others its just a little faster.
		 */
543
		free_pgd_range(tlb, old_start, old_end, new_end,
544
			vma->vm_next ? vma->vm_next->vm_start : 0);
L
Linus Torvalds 已提交
545
	}
546 547 548 549 550 551 552 553
	tlb_finish_mmu(tlb, new_end, old_end);

	/*
	 * shrink the vma to just the new range.
	 */
	vma_adjust(vma, new_start, new_end, vma->vm_pgoff, NULL);

	return 0;
L
Linus Torvalds 已提交
554 555 556 557
}

#define EXTRA_STACK_VM_PAGES	20	/* random */

558 559 560 561
/*
 * Finalizes the stack vm_area_struct. The flags and permissions are updated,
 * the stack is optionally relocated, and some extra space is added.
 */
L
Linus Torvalds 已提交
562 563 564 565
int setup_arg_pages(struct linux_binprm *bprm,
		    unsigned long stack_top,
		    int executable_stack)
{
566 567
	unsigned long ret;
	unsigned long stack_shift;
L
Linus Torvalds 已提交
568
	struct mm_struct *mm = current->mm;
569 570 571 572
	struct vm_area_struct *vma = bprm->vma;
	struct vm_area_struct *prev = NULL;
	unsigned long vm_flags;
	unsigned long stack_base;
L
Linus Torvalds 已提交
573 574 575 576 577 578 579

#ifdef CONFIG_STACK_GROWSUP
	/* Limit stack size to 1GB */
	stack_base = current->signal->rlim[RLIMIT_STACK].rlim_max;
	if (stack_base > (1 << 30))
		stack_base = 1 << 30;

580 581 582
	/* Make sure we didn't let the argument array grow too large. */
	if (vma->vm_end - vma->vm_start > stack_base)
		return -ENOMEM;
L
Linus Torvalds 已提交
583

584
	stack_base = PAGE_ALIGN(stack_top - stack_base);
L
Linus Torvalds 已提交
585

586 587 588
	stack_shift = vma->vm_start - stack_base;
	mm->arg_start = bprm->p - stack_shift;
	bprm->p = vma->vm_end - stack_shift;
L
Linus Torvalds 已提交
589
#else
590 591 592 593 594
	stack_top = arch_align_stack(stack_top);
	stack_top = PAGE_ALIGN(stack_top);
	stack_shift = vma->vm_end - stack_top;

	bprm->p -= stack_shift;
L
Linus Torvalds 已提交
595 596 597 598
	mm->arg_start = bprm->p;
#endif

	if (bprm->loader)
599 600
		bprm->loader -= stack_shift;
	bprm->exec -= stack_shift;
L
Linus Torvalds 已提交
601 602

	down_write(&mm->mmap_sem);
603
	vm_flags = VM_STACK_FLAGS;
604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625

	/*
	 * Adjust stack execute permissions; explicitly enable for
	 * EXSTACK_ENABLE_X, disable for EXSTACK_DISABLE_X and leave alone
	 * (arch default) otherwise.
	 */
	if (unlikely(executable_stack == EXSTACK_ENABLE_X))
		vm_flags |= VM_EXEC;
	else if (executable_stack == EXSTACK_DISABLE_X)
		vm_flags &= ~VM_EXEC;
	vm_flags |= mm->def_flags;

	ret = mprotect_fixup(vma, &prev, vma->vm_start, vma->vm_end,
			vm_flags);
	if (ret)
		goto out_unlock;
	BUG_ON(prev != vma);

	/* Move stack pages down in memory. */
	if (stack_shift) {
		ret = shift_arg_pages(vma, stack_shift);
		if (ret) {
L
Linus Torvalds 已提交
626 627 628 629 630
			up_write(&mm->mmap_sem);
			return ret;
		}
	}

631 632 633 634 635 636 637 638 639 640
#ifdef CONFIG_STACK_GROWSUP
	stack_base = vma->vm_end + EXTRA_STACK_VM_PAGES * PAGE_SIZE;
#else
	stack_base = vma->vm_start - EXTRA_STACK_VM_PAGES * PAGE_SIZE;
#endif
	ret = expand_stack(vma, stack_base);
	if (ret)
		ret = -EFAULT;

out_unlock:
L
Linus Torvalds 已提交
641 642 643 644 645 646 647 648 649 650
	up_write(&mm->mmap_sem);
	return 0;
}
EXPORT_SYMBOL(setup_arg_pages);

#endif /* CONFIG_MMU */

struct file *open_exec(const char *name)
{
	struct file *file;
651
	int err;
L
Linus Torvalds 已提交
652

653 654 655 656
	file = do_filp_open(AT_FDCWD, name,
				O_LARGEFILE | O_RDONLY | FMODE_EXEC, 0,
				MAY_EXEC | MAY_OPEN);
	if (IS_ERR(file))
657 658 659
		goto out;

	err = -EACCES;
660 661
	if (!S_ISREG(file->f_path.dentry->d_inode->i_mode))
		goto exit;
662

663 664
	if (file->f_path.mnt->mnt_flags & MNT_NOEXEC)
		goto exit;
665

666 667
	fsnotify_open(file->f_path.dentry);

668
	err = deny_write_access(file);
669 670
	if (err)
		goto exit;
L
Linus Torvalds 已提交
671

672
out:
673 674
	return file;

675 676
exit:
	fput(file);
677 678
	return ERR_PTR(err);
}
L
Linus Torvalds 已提交
679 680
EXPORT_SYMBOL(open_exec);

M
Mimi Zohar 已提交
681 682
int kernel_read(struct file *file, loff_t offset,
		char *addr, unsigned long count)
L
Linus Torvalds 已提交
683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712
{
	mm_segment_t old_fs;
	loff_t pos = offset;
	int result;

	old_fs = get_fs();
	set_fs(get_ds());
	/* The cast to a user pointer is valid due to the set_fs() */
	result = vfs_read(file, (void __user *)addr, count, &pos);
	set_fs(old_fs);
	return result;
}

EXPORT_SYMBOL(kernel_read);

static int exec_mmap(struct mm_struct *mm)
{
	struct task_struct *tsk;
	struct mm_struct * old_mm, *active_mm;

	/* Notify parent that we're no longer interested in the old VM */
	tsk = current;
	old_mm = current->mm;
	mm_release(tsk, old_mm);

	if (old_mm) {
		/*
		 * Make sure that if there is a core dump in progress
		 * for the old mm, we get out and die instead of going
		 * through with the exec.  We must hold mmap_sem around
713
		 * checking core_state and changing tsk->mm.
L
Linus Torvalds 已提交
714 715
		 */
		down_read(&old_mm->mmap_sem);
716
		if (unlikely(old_mm->core_state)) {
L
Linus Torvalds 已提交
717 718 719 720 721 722 723 724 725 726 727 728 729
			up_read(&old_mm->mmap_sem);
			return -EINTR;
		}
	}
	task_lock(tsk);
	active_mm = tsk->active_mm;
	tsk->mm = mm;
	tsk->active_mm = mm;
	activate_mm(active_mm, mm);
	task_unlock(tsk);
	arch_pick_mmap_layout(mm);
	if (old_mm) {
		up_read(&old_mm->mmap_sem);
730
		BUG_ON(active_mm != old_mm);
731
		mm_update_next_owner(old_mm);
L
Linus Torvalds 已提交
732 733 734 735 736 737 738 739 740 741 742 743 744
		mmput(old_mm);
		return 0;
	}
	mmdrop(active_mm);
	return 0;
}

/*
 * This function makes sure the current process has its own signal table,
 * so that flush_signal_handlers can later reset the handlers without
 * disturbing other processes.  (Other processes might share the signal
 * table via the CLONE_SIGHAND option to clone().)
 */
745
static int de_thread(struct task_struct *tsk)
L
Linus Torvalds 已提交
746 747
{
	struct signal_struct *sig = tsk->signal;
748
	struct sighand_struct *oldsighand = tsk->sighand;
L
Linus Torvalds 已提交
749 750 751
	spinlock_t *lock = &oldsighand->siglock;
	int count;

752
	if (thread_group_empty(tsk))
L
Linus Torvalds 已提交
753 754 755 756 757 758
		goto no_thread_group;

	/*
	 * Kill all other threads in the thread group.
	 */
	spin_lock_irq(lock);
759
	if (signal_group_exit(sig)) {
L
Linus Torvalds 已提交
760 761 762 763 764 765 766
		/*
		 * Another group action in progress, just
		 * return so that the signal is processed.
		 */
		spin_unlock_irq(lock);
		return -EAGAIN;
	}
767
	sig->group_exit_task = tsk;
768
	zap_other_threads(tsk);
L
Linus Torvalds 已提交
769

770 771
	/* Account for the thread group leader hanging around: */
	count = thread_group_leader(tsk) ? 1 : 2;
772
	sig->notify_count = count;
L
Linus Torvalds 已提交
773 774 775 776 777 778 779 780 781 782 783 784 785
	while (atomic_read(&sig->count) > count) {
		__set_current_state(TASK_UNINTERRUPTIBLE);
		spin_unlock_irq(lock);
		schedule();
		spin_lock_irq(lock);
	}
	spin_unlock_irq(lock);

	/*
	 * At this point all other threads have exited, all we have to
	 * do is to wait for the thread group leader to become inactive,
	 * and to assume its PID:
	 */
786
	if (!thread_group_leader(tsk)) {
787
		struct task_struct *leader = tsk->group_leader;
788

789
		sig->notify_count = -1;	/* for exit_notify() */
790 791 792 793 794 795 796 797
		for (;;) {
			write_lock_irq(&tasklist_lock);
			if (likely(leader->exit_state))
				break;
			__set_current_state(TASK_UNINTERRUPTIBLE);
			write_unlock_irq(&tasklist_lock);
			schedule();
		}
L
Linus Torvalds 已提交
798

799 800 801 802 803 804 805 806 807 808
		/*
		 * The only record we have of the real-time age of a
		 * process, regardless of execs it's done, is start_time.
		 * All the past CPU time is accumulated in signal_struct
		 * from sister threads now dead.  But in this non-leader
		 * exec, nothing survives from the original leader thread,
		 * whose birth marks the true age of this process now.
		 * When we take on its identity by switching to its PID, we
		 * also take its birthdate (always earlier than our own).
		 */
809
		tsk->start_time = leader->start_time;
810

811 812
		BUG_ON(!same_thread_group(leader, tsk));
		BUG_ON(has_group_leader_pid(tsk));
L
Linus Torvalds 已提交
813 814 815 816 817 818
		/*
		 * An exec() starts a new thread group with the
		 * TGID of the previous thread group. Rehash the
		 * two threads with a switched PID, and release
		 * the former thread group leader:
		 */
819 820

		/* Become a process group leader with the old leader's pid.
821 822
		 * The old leader becomes a thread of the this thread group.
		 * Note: The old leader also uses this pid until release_task
823 824
		 *       is called.  Odd but simple and correct.
		 */
825 826
		detach_pid(tsk, PIDTYPE_PID);
		tsk->pid = leader->pid;
827
		attach_pid(tsk, PIDTYPE_PID,  task_pid(leader));
828 829 830
		transfer_pid(leader, tsk, PIDTYPE_PGID);
		transfer_pid(leader, tsk, PIDTYPE_SID);
		list_replace_rcu(&leader->tasks, &tsk->tasks);
L
Linus Torvalds 已提交
831

832 833
		tsk->group_leader = tsk;
		leader->group_leader = tsk;
834

835
		tsk->exit_signal = SIGCHLD;
836 837 838

		BUG_ON(leader->exit_state != EXIT_ZOMBIE);
		leader->exit_state = EXIT_DEAD;
L
Linus Torvalds 已提交
839
		write_unlock_irq(&tasklist_lock);
840 841

		release_task(leader);
842
	}
L
Linus Torvalds 已提交
843

844 845
	sig->group_exit_task = NULL;
	sig->notify_count = 0;
L
Linus Torvalds 已提交
846 847 848

no_thread_group:
	exit_itimers(sig);
849
	flush_itimer_signals();
850

851 852
	if (atomic_read(&oldsighand->count) != 1) {
		struct sighand_struct *newsighand;
L
Linus Torvalds 已提交
853
		/*
854 855
		 * This ->sighand is shared with the CLONE_SIGHAND
		 * but not CLONE_THREAD task, switch to the new one.
L
Linus Torvalds 已提交
856
		 */
857 858 859 860
		newsighand = kmem_cache_alloc(sighand_cachep, GFP_KERNEL);
		if (!newsighand)
			return -ENOMEM;

L
Linus Torvalds 已提交
861 862 863 864 865 866
		atomic_set(&newsighand->count, 1);
		memcpy(newsighand->action, oldsighand->action,
		       sizeof(newsighand->action));

		write_lock_irq(&tasklist_lock);
		spin_lock(&oldsighand->siglock);
867
		rcu_assign_pointer(tsk->sighand, newsighand);
L
Linus Torvalds 已提交
868 869 870
		spin_unlock(&oldsighand->siglock);
		write_unlock_irq(&tasklist_lock);

871
		__cleanup_sighand(oldsighand);
L
Linus Torvalds 已提交
872 873
	}

874
	BUG_ON(!thread_group_leader(tsk));
L
Linus Torvalds 已提交
875 876
	return 0;
}
O
Oleg Nesterov 已提交
877

L
Linus Torvalds 已提交
878 879 880 881
/*
 * These functions flushes out all traces of the currently running executable
 * so that a new one can be started
 */
882
static void flush_old_files(struct files_struct * files)
L
Linus Torvalds 已提交
883 884
{
	long j = -1;
885
	struct fdtable *fdt;
L
Linus Torvalds 已提交
886 887 888 889 890 891 892

	spin_lock(&files->file_lock);
	for (;;) {
		unsigned long set, i;

		j++;
		i = j * __NFDBITS;
893
		fdt = files_fdtable(files);
894
		if (i >= fdt->max_fds)
L
Linus Torvalds 已提交
895
			break;
896
		set = fdt->close_on_exec->fds_bits[j];
L
Linus Torvalds 已提交
897 898
		if (!set)
			continue;
899
		fdt->close_on_exec->fds_bits[j] = 0;
L
Linus Torvalds 已提交
900 901 902 903 904 905 906 907 908 909 910 911
		spin_unlock(&files->file_lock);
		for ( ; set ; i++,set >>= 1) {
			if (set & 1) {
				sys_close(i);
			}
		}
		spin_lock(&files->file_lock);

	}
	spin_unlock(&files->file_lock);
}

912
char *get_task_comm(char *buf, struct task_struct *tsk)
L
Linus Torvalds 已提交
913 914 915 916 917
{
	/* buf must be at least sizeof(tsk->comm) in size */
	task_lock(tsk);
	strncpy(buf, tsk->comm, sizeof(tsk->comm));
	task_unlock(tsk);
918
	return buf;
L
Linus Torvalds 已提交
919 920 921 922 923 924 925
}

void set_task_comm(struct task_struct *tsk, char *buf)
{
	task_lock(tsk);
	strlcpy(tsk->comm, buf, sizeof(tsk->comm));
	task_unlock(tsk);
926
	perf_event_comm(tsk);
L
Linus Torvalds 已提交
927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942
}

int flush_old_exec(struct linux_binprm * bprm)
{
	char * name;
	int i, ch, retval;
	char tcomm[sizeof(current->comm)];

	/*
	 * Make sure we have a private signal table and that
	 * we are unassociated from the previous thread group.
	 */
	retval = de_thread(current);
	if (retval)
		goto out;

M
Matt Helsley 已提交
943 944
	set_mm_exe_file(bprm->mm, bprm->file);

L
Linus Torvalds 已提交
945 946 947 948 949
	/*
	 * Release all of the old mmap stuff
	 */
	retval = exec_mmap(bprm->mm);
	if (retval)
950
		goto out;
L
Linus Torvalds 已提交
951 952 953 954 955 956

	bprm->mm = NULL;		/* We're using it now */

	/* This is the point of no return */
	current->sas_ss_sp = current->sas_ss_size = 0;

957
	if (current_euid() == current_uid() && current_egid() == current_gid())
958
		set_dumpable(current->mm, 1);
A
Alan Cox 已提交
959
	else
960
		set_dumpable(current->mm, suid_dumpable);
A
Alan Cox 已提交
961

L
Linus Torvalds 已提交
962
	name = bprm->filename;
963 964

	/* Copies the binary name from after last slash */
L
Linus Torvalds 已提交
965 966
	for (i=0; (ch = *(name++)) != '\0';) {
		if (ch == '/')
967
			i = 0; /* overwrite what we wrote */
L
Linus Torvalds 已提交
968 969 970 971 972 973 974 975 976 977
		else
			if (i < (sizeof(tcomm) - 1))
				tcomm[i++] = ch;
	}
	tcomm[i] = '\0';
	set_task_comm(current, tcomm);

	current->flags &= ~PF_RANDOMIZE;
	flush_thread();

978 979 980 981 982 983
	/* Set the new mm task size. We have to do that late because it may
	 * depend on TIF_32BIT which is only updated in flush_thread() on
	 * some architectures like powerpc
	 */
	current->mm->task_size = TASK_SIZE;

984 985 986
	/* install the new credentials */
	if (bprm->cred->uid != current_euid() ||
	    bprm->cred->gid != current_egid()) {
987 988
		current->pdeath_signal = 0;
	} else if (file_permission(bprm->file, MAY_READ) ||
989
		   bprm->interp_flags & BINPRM_FLAGS_ENFORCE_NONDUMP) {
990
		set_dumpable(current->mm, suid_dumpable);
L
Linus Torvalds 已提交
991 992
	}

993 994
	current->personality &= ~bprm->per_clear;

I
Ingo Molnar 已提交
995 996 997 998 999
	/*
	 * Flush performance counters when crossing a
	 * security domain:
	 */
	if (!get_dumpable(current->mm))
1000
		perf_event_exit_task(current);
I
Ingo Molnar 已提交
1001

L
Linus Torvalds 已提交
1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017
	/* An exec changes our domain. We are no longer part of the thread
	   group */

	current->self_exec_id++;
			
	flush_signal_handlers(current, 0);
	flush_old_files(current->files);

	return 0;

out:
	return retval;
}

EXPORT_SYMBOL(flush_old_exec);

1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046
/*
 * Prepare credentials and lock ->cred_guard_mutex.
 * install_exec_creds() commits the new creds and drops the lock.
 * Or, if exec fails before, free_bprm() should release ->cred and
 * and unlock.
 */
int prepare_bprm_creds(struct linux_binprm *bprm)
{
	if (mutex_lock_interruptible(&current->cred_guard_mutex))
		return -ERESTARTNOINTR;

	bprm->cred = prepare_exec_creds();
	if (likely(bprm->cred))
		return 0;

	mutex_unlock(&current->cred_guard_mutex);
	return -ENOMEM;
}

void free_bprm(struct linux_binprm *bprm)
{
	free_arg_pages(bprm);
	if (bprm->cred) {
		mutex_unlock(&current->cred_guard_mutex);
		abort_creds(bprm->cred);
	}
	kfree(bprm);
}

1047 1048 1049 1050 1051 1052 1053 1054 1055
/*
 * install the new credentials for this executable
 */
void install_exec_creds(struct linux_binprm *bprm)
{
	security_bprm_committing_creds(bprm);

	commit_creds(bprm->cred);
	bprm->cred = NULL;
1056 1057
	/*
	 * cred_guard_mutex must be held at least to this point to prevent
1058
	 * ptrace_attach() from altering our determination of the task's
1059 1060
	 * credentials; any time after this it may be unlocked.
	 */
1061
	security_bprm_committed_creds(bprm);
1062
	mutex_unlock(&current->cred_guard_mutex);
1063 1064 1065 1066 1067
}
EXPORT_SYMBOL(install_exec_creds);

/*
 * determine how safe it is to execute the proposed program
1068
 * - the caller must hold current->cred_guard_mutex to protect against
1069 1070
 *   PTRACE_ATTACH
 */
A
Al Viro 已提交
1071
int check_unsafe_exec(struct linux_binprm *bprm)
1072
{
D
David Howells 已提交
1073
	struct task_struct *p = current, *t;
1074
	unsigned n_fs;
A
Al Viro 已提交
1075
	int res = 0;
1076 1077 1078

	bprm->unsafe = tracehook_unsafe_exec(p);

D
David Howells 已提交
1079
	n_fs = 1;
A
Al Viro 已提交
1080
	write_lock(&p->fs->lock);
1081
	rcu_read_lock();
D
David Howells 已提交
1082 1083 1084 1085
	for (t = next_thread(p); t != p; t = next_thread(t)) {
		if (t->fs == p->fs)
			n_fs++;
	}
1086
	rcu_read_unlock();
D
David Howells 已提交
1087

1088
	if (p->fs->users > n_fs) {
1089
		bprm->unsafe |= LSM_UNSAFE_SHARE;
A
Al Viro 已提交
1090
	} else {
1091 1092 1093 1094 1095
		res = -EAGAIN;
		if (!p->fs->in_exec) {
			p->fs->in_exec = 1;
			res = 1;
		}
A
Al Viro 已提交
1096 1097 1098 1099
	}
	write_unlock(&p->fs->lock);

	return res;
1100 1101
}

L
Linus Torvalds 已提交
1102 1103 1104
/* 
 * Fill the binprm structure from the inode. 
 * Check permissions, then read the first 128 (BINPRM_BUF_SIZE) bytes
1105 1106
 *
 * This may be called multiple times for binary chains (scripts for example).
L
Linus Torvalds 已提交
1107 1108 1109
 */
int prepare_binprm(struct linux_binprm *bprm)
{
1110
	umode_t mode;
1111
	struct inode * inode = bprm->file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1112 1113 1114 1115 1116 1117
	int retval;

	mode = inode->i_mode;
	if (bprm->file->f_op == NULL)
		return -EACCES;

1118 1119 1120
	/* clear any previous set[ug]id data from a previous binary */
	bprm->cred->euid = current_euid();
	bprm->cred->egid = current_egid();
L
Linus Torvalds 已提交
1121

1122
	if (!(bprm->file->f_path.mnt->mnt_flags & MNT_NOSUID)) {
L
Linus Torvalds 已提交
1123 1124
		/* Set-uid? */
		if (mode & S_ISUID) {
1125 1126
			bprm->per_clear |= PER_CLEAR_ON_SETID;
			bprm->cred->euid = inode->i_uid;
L
Linus Torvalds 已提交
1127 1128 1129 1130 1131 1132 1133 1134 1135
		}

		/* Set-gid? */
		/*
		 * If setgid is set but no group execute bit then this
		 * is a candidate for mandatory locking, not a setgid
		 * executable.
		 */
		if ((mode & (S_ISGID | S_IXGRP)) == (S_ISGID | S_IXGRP)) {
1136 1137
			bprm->per_clear |= PER_CLEAR_ON_SETID;
			bprm->cred->egid = inode->i_gid;
L
Linus Torvalds 已提交
1138 1139 1140 1141
		}
	}

	/* fill in binprm security blob */
1142
	retval = security_bprm_set_creds(bprm);
L
Linus Torvalds 已提交
1143 1144
	if (retval)
		return retval;
1145
	bprm->cred_prepared = 1;
L
Linus Torvalds 已提交
1146

1147 1148
	memset(bprm->buf, 0, BINPRM_BUF_SIZE);
	return kernel_read(bprm->file, 0, bprm->buf, BINPRM_BUF_SIZE);
L
Linus Torvalds 已提交
1149 1150 1151 1152
}

EXPORT_SYMBOL(prepare_binprm);

N
Nick Piggin 已提交
1153 1154 1155 1156 1157
/*
 * Arguments are '\0' separated strings found at the location bprm->p
 * points to; chop off the first by relocating brpm->p to right after
 * the first '\0' encountered.
 */
1158
int remove_arg_zero(struct linux_binprm *bprm)
L
Linus Torvalds 已提交
1159
{
1160 1161 1162 1163
	int ret = 0;
	unsigned long offset;
	char *kaddr;
	struct page *page;
N
Nick Piggin 已提交
1164

1165 1166
	if (!bprm->argc)
		return 0;
L
Linus Torvalds 已提交
1167

1168 1169 1170 1171 1172 1173 1174 1175
	do {
		offset = bprm->p & ~PAGE_MASK;
		page = get_arg_page(bprm, bprm->p, 0);
		if (!page) {
			ret = -EFAULT;
			goto out;
		}
		kaddr = kmap_atomic(page, KM_USER0);
N
Nick Piggin 已提交
1176

1177 1178 1179
		for (; offset < PAGE_SIZE && kaddr[offset];
				offset++, bprm->p++)
			;
N
Nick Piggin 已提交
1180

1181 1182
		kunmap_atomic(kaddr, KM_USER0);
		put_arg_page(page);
N
Nick Piggin 已提交
1183

1184 1185 1186
		if (offset == PAGE_SIZE)
			free_arg_page(bprm, (bprm->p >> PAGE_SHIFT) - 1);
	} while (offset == PAGE_SIZE);
N
Nick Piggin 已提交
1187

1188 1189 1190
	bprm->p++;
	bprm->argc--;
	ret = 0;
N
Nick Piggin 已提交
1191

1192 1193
out:
	return ret;
L
Linus Torvalds 已提交
1194 1195 1196 1197 1198 1199 1200 1201
}
EXPORT_SYMBOL(remove_arg_zero);

/*
 * cycle the list of binary formats handler, until one recognizes the image
 */
int search_binary_handler(struct linux_binprm *bprm,struct pt_regs *regs)
{
1202
	unsigned int depth = bprm->recursion_depth;
L
Linus Torvalds 已提交
1203 1204 1205 1206
	int try,retval;
	struct linux_binfmt *fmt;

	retval = security_bprm_check(bprm);
M
Mimi Zohar 已提交
1207 1208 1209
	if (retval)
		return retval;
	retval = ima_bprm_check(bprm);
L
Linus Torvalds 已提交
1210 1211 1212 1213 1214 1215
	if (retval)
		return retval;

	/* kernel module loader fixup */
	/* so we don't try to load run modprobe in kernel space. */
	set_fs(USER_DS);
A
Al Viro 已提交
1216 1217 1218 1219 1220

	retval = audit_bprm(bprm);
	if (retval)
		return retval;

L
Linus Torvalds 已提交
1221 1222 1223
	retval = -ENOENT;
	for (try=0; try<2; try++) {
		read_lock(&binfmt_lock);
A
Alexey Dobriyan 已提交
1224
		list_for_each_entry(fmt, &formats, lh) {
L
Linus Torvalds 已提交
1225 1226 1227 1228 1229 1230 1231
			int (*fn)(struct linux_binprm *, struct pt_regs *) = fmt->load_binary;
			if (!fn)
				continue;
			if (!try_module_get(fmt->module))
				continue;
			read_unlock(&binfmt_lock);
			retval = fn(bprm, regs);
1232 1233 1234 1235 1236 1237
			/*
			 * Restore the depth counter to its starting value
			 * in this call, so we don't have to rely on every
			 * load_binary function to restore it on return.
			 */
			bprm->recursion_depth = depth;
L
Linus Torvalds 已提交
1238
			if (retval >= 0) {
1239 1240
				if (depth == 0)
					tracehook_report_exec(fmt, bprm, regs);
L
Linus Torvalds 已提交
1241 1242 1243 1244 1245 1246
				put_binfmt(fmt);
				allow_write_access(bprm->file);
				if (bprm->file)
					fput(bprm->file);
				bprm->file = NULL;
				current->did_exec = 1;
M
Matt Helsley 已提交
1247
				proc_exec_connector(current);
L
Linus Torvalds 已提交
1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261
				return retval;
			}
			read_lock(&binfmt_lock);
			put_binfmt(fmt);
			if (retval != -ENOEXEC || bprm->mm == NULL)
				break;
			if (!bprm->file) {
				read_unlock(&binfmt_lock);
				return retval;
			}
		}
		read_unlock(&binfmt_lock);
		if (retval != -ENOEXEC || bprm->mm == NULL) {
			break;
J
Johannes Berg 已提交
1262 1263
#ifdef CONFIG_MODULES
		} else {
L
Linus Torvalds 已提交
1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288
#define printable(c) (((c)=='\t') || ((c)=='\n') || (0x20<=(c) && (c)<=0x7e))
			if (printable(bprm->buf[0]) &&
			    printable(bprm->buf[1]) &&
			    printable(bprm->buf[2]) &&
			    printable(bprm->buf[3]))
				break; /* -ENOEXEC */
			request_module("binfmt-%04x", *(unsigned short *)(&bprm->buf[2]));
#endif
		}
	}
	return retval;
}

EXPORT_SYMBOL(search_binary_handler);

/*
 * sys_execve() executes a new program.
 */
int do_execve(char * filename,
	char __user *__user *argv,
	char __user *__user *envp,
	struct pt_regs * regs)
{
	struct linux_binprm *bprm;
	struct file *file;
1289
	struct files_struct *displaced;
1290
	bool clear_in_exec;
L
Linus Torvalds 已提交
1291 1292
	int retval;

1293
	retval = unshare_files(&displaced);
1294 1295 1296
	if (retval)
		goto out_ret;

L
Linus Torvalds 已提交
1297
	retval = -ENOMEM;
1298
	bprm = kzalloc(sizeof(*bprm), GFP_KERNEL);
L
Linus Torvalds 已提交
1299
	if (!bprm)
1300
		goto out_files;
L
Linus Torvalds 已提交
1301

1302 1303
	retval = prepare_bprm_creds(bprm);
	if (retval)
1304
		goto out_free;
A
Al Viro 已提交
1305 1306

	retval = check_unsafe_exec(bprm);
1307
	if (retval < 0)
1308
		goto out_free;
1309
	clear_in_exec = retval;
1310
	current->in_execve = 1;
1311

L
Linus Torvalds 已提交
1312 1313 1314
	file = open_exec(filename);
	retval = PTR_ERR(file);
	if (IS_ERR(file))
A
Al Viro 已提交
1315
		goto out_unmark;
L
Linus Torvalds 已提交
1316 1317 1318 1319 1320 1321 1322

	sched_exec();

	bprm->file = file;
	bprm->filename = filename;
	bprm->interp = filename;

1323 1324 1325
	retval = bprm_mm_init(bprm);
	if (retval)
		goto out_file;
L
Linus Torvalds 已提交
1326

1327
	bprm->argc = count(argv, MAX_ARG_STRINGS);
L
Linus Torvalds 已提交
1328
	if ((retval = bprm->argc) < 0)
1329
		goto out;
L
Linus Torvalds 已提交
1330

1331
	bprm->envc = count(envp, MAX_ARG_STRINGS);
L
Linus Torvalds 已提交
1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351
	if ((retval = bprm->envc) < 0)
		goto out;

	retval = prepare_binprm(bprm);
	if (retval < 0)
		goto out;

	retval = copy_strings_kernel(1, &bprm->filename, bprm);
	if (retval < 0)
		goto out;

	bprm->exec = bprm->p;
	retval = copy_strings(bprm->envc, envp, bprm);
	if (retval < 0)
		goto out;

	retval = copy_strings(bprm->argc, argv, bprm);
	if (retval < 0)
		goto out;

O
Oleg Nesterov 已提交
1352
	current->flags &= ~PF_KTHREAD;
L
Linus Torvalds 已提交
1353
	retval = search_binary_handler(bprm,regs);
1354 1355
	if (retval < 0)
		goto out;
L
Linus Torvalds 已提交
1356

1357
	/* execve succeeded */
A
Al Viro 已提交
1358
	current->fs->in_exec = 0;
1359
	current->in_execve = 0;
1360 1361 1362 1363 1364
	acct_update_integrals(current);
	free_bprm(bprm);
	if (displaced)
		put_files_struct(displaced);
	return retval;
L
Linus Torvalds 已提交
1365

1366
out:
L
Linus Torvalds 已提交
1367
	if (bprm->mm)
1368
		mmput (bprm->mm);
L
Linus Torvalds 已提交
1369 1370 1371 1372 1373 1374

out_file:
	if (bprm->file) {
		allow_write_access(bprm->file);
		fput(bprm->file);
	}
1375

A
Al Viro 已提交
1376
out_unmark:
1377 1378
	if (clear_in_exec)
		current->fs->in_exec = 0;
1379
	current->in_execve = 0;
1380 1381

out_free:
1382
	free_bprm(bprm);
L
Linus Torvalds 已提交
1383

1384
out_files:
1385 1386
	if (displaced)
		reset_files_struct(displaced);
L
Linus Torvalds 已提交
1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410
out_ret:
	return retval;
}

int set_binfmt(struct linux_binfmt *new)
{
	struct linux_binfmt *old = current->binfmt;

	if (new) {
		if (!try_module_get(new->module))
			return -1;
	}
	current->binfmt = new;
	if (old)
		module_put(old->module);
	return 0;
}

EXPORT_SYMBOL(set_binfmt);

/* format_corename will inspect the pattern parameter, and output a
 * name into corename, which must have space for at least
 * CORENAME_MAX_SIZE bytes plus one byte for the zero terminator.
 */
1411
static int format_corename(char *corename, long signr)
L
Linus Torvalds 已提交
1412
{
1413
	const struct cred *cred = current_cred();
1414 1415
	const char *pat_ptr = core_pattern;
	int ispipe = (*pat_ptr == '|');
L
Linus Torvalds 已提交
1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441
	char *out_ptr = corename;
	char *const out_end = corename + CORENAME_MAX_SIZE;
	int rc;
	int pid_in_pattern = 0;

	/* Repeat as long as we have more pattern to process and more output
	   space */
	while (*pat_ptr) {
		if (*pat_ptr != '%') {
			if (out_ptr == out_end)
				goto out;
			*out_ptr++ = *pat_ptr++;
		} else {
			switch (*++pat_ptr) {
			case 0:
				goto out;
			/* Double percent, output one percent */
			case '%':
				if (out_ptr == out_end)
					goto out;
				*out_ptr++ = '%';
				break;
			/* pid */
			case 'p':
				pid_in_pattern = 1;
				rc = snprintf(out_ptr, out_end - out_ptr,
1442
					      "%d", task_tgid_vnr(current));
L
Linus Torvalds 已提交
1443 1444 1445 1446 1447 1448 1449
				if (rc > out_end - out_ptr)
					goto out;
				out_ptr += rc;
				break;
			/* uid */
			case 'u':
				rc = snprintf(out_ptr, out_end - out_ptr,
1450
					      "%d", cred->uid);
L
Linus Torvalds 已提交
1451 1452 1453 1454 1455 1456 1457
				if (rc > out_end - out_ptr)
					goto out;
				out_ptr += rc;
				break;
			/* gid */
			case 'g':
				rc = snprintf(out_ptr, out_end - out_ptr,
1458
					      "%d", cred->gid);
L
Linus Torvalds 已提交
1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485
				if (rc > out_end - out_ptr)
					goto out;
				out_ptr += rc;
				break;
			/* signal that caused the coredump */
			case 's':
				rc = snprintf(out_ptr, out_end - out_ptr,
					      "%ld", signr);
				if (rc > out_end - out_ptr)
					goto out;
				out_ptr += rc;
				break;
			/* UNIX time of coredump */
			case 't': {
				struct timeval tv;
				do_gettimeofday(&tv);
				rc = snprintf(out_ptr, out_end - out_ptr,
					      "%lu", tv.tv_sec);
				if (rc > out_end - out_ptr)
					goto out;
				out_ptr += rc;
				break;
			}
			/* hostname */
			case 'h':
				down_read(&uts_sem);
				rc = snprintf(out_ptr, out_end - out_ptr,
1486
					      "%s", utsname()->nodename);
L
Linus Torvalds 已提交
1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499
				up_read(&uts_sem);
				if (rc > out_end - out_ptr)
					goto out;
				out_ptr += rc;
				break;
			/* executable */
			case 'e':
				rc = snprintf(out_ptr, out_end - out_ptr,
					      "%s", current->comm);
				if (rc > out_end - out_ptr)
					goto out;
				out_ptr += rc;
				break;
1500 1501 1502 1503 1504 1505 1506 1507
			/* core limit size */
			case 'c':
				rc = snprintf(out_ptr, out_end - out_ptr,
					      "%lu", current->signal->rlim[RLIMIT_CORE].rlim_cur);
				if (rc > out_end - out_ptr)
					goto out;
				out_ptr += rc;
				break;
L
Linus Torvalds 已提交
1508 1509 1510 1511 1512 1513 1514 1515 1516 1517
			default:
				break;
			}
			++pat_ptr;
		}
	}
	/* Backward compatibility with core_uses_pid:
	 *
	 * If core_pattern does not include a %p (as is the default)
	 * and core_uses_pid is set, then .%pid will be appended to
1518
	 * the filename. Do not do this for piped commands. */
1519
	if (!ispipe && !pid_in_pattern && core_uses_pid) {
L
Linus Torvalds 已提交
1520
		rc = snprintf(out_ptr, out_end - out_ptr,
1521
			      ".%d", task_tgid_vnr(current));
L
Linus Torvalds 已提交
1522 1523 1524 1525
		if (rc > out_end - out_ptr)
			goto out;
		out_ptr += rc;
	}
1526
out:
L
Linus Torvalds 已提交
1527
	*out_ptr = 0;
1528
	return ispipe;
L
Linus Torvalds 已提交
1529 1530
}

1531
static int zap_process(struct task_struct *start)
1532 1533
{
	struct task_struct *t;
1534
	int nr = 0;
1535

1536 1537
	start->signal->flags = SIGNAL_GROUP_EXIT;
	start->signal->group_stop_count = 0;
1538 1539 1540 1541

	t = start;
	do {
		if (t != current && t->mm) {
1542 1543
			sigaddset(&t->pending.signal, SIGKILL);
			signal_wake_up(t, 1);
1544
			nr++;
1545
		}
1546
	} while_each_thread(start, t);
1547 1548

	return nr;
1549 1550
}

1551
static inline int zap_threads(struct task_struct *tsk, struct mm_struct *mm,
1552
				struct core_state *core_state, int exit_code)
L
Linus Torvalds 已提交
1553 1554
{
	struct task_struct *g, *p;
1555
	unsigned long flags;
1556
	int nr = -EAGAIN;
1557 1558

	spin_lock_irq(&tsk->sighand->siglock);
1559
	if (!signal_group_exit(tsk->signal)) {
1560
		mm->core_state = core_state;
1561
		tsk->signal->group_exit_code = exit_code;
1562
		nr = zap_process(tsk);
L
Linus Torvalds 已提交
1563
	}
1564
	spin_unlock_irq(&tsk->sighand->siglock);
1565 1566
	if (unlikely(nr < 0))
		return nr;
L
Linus Torvalds 已提交
1567

1568
	if (atomic_read(&mm->mm_users) == nr + 1)
1569
		goto done;
1570 1571
	/*
	 * We should find and kill all tasks which use this mm, and we should
1572
	 * count them correctly into ->nr_threads. We don't take tasklist
1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599
	 * lock, but this is safe wrt:
	 *
	 * fork:
	 *	None of sub-threads can fork after zap_process(leader). All
	 *	processes which were created before this point should be
	 *	visible to zap_threads() because copy_process() adds the new
	 *	process to the tail of init_task.tasks list, and lock/unlock
	 *	of ->siglock provides a memory barrier.
	 *
	 * do_exit:
	 *	The caller holds mm->mmap_sem. This means that the task which
	 *	uses this mm can't pass exit_mm(), so it can't exit or clear
	 *	its ->mm.
	 *
	 * de_thread:
	 *	It does list_replace_rcu(&leader->tasks, &current->tasks),
	 *	we must see either old or new leader, this does not matter.
	 *	However, it can change p->sighand, so lock_task_sighand(p)
	 *	must be used. Since p->mm != NULL and we hold ->mmap_sem
	 *	it can't fail.
	 *
	 *	Note also that "g" can be the old leader with ->mm == NULL
	 *	and already unhashed and thus removed from ->thread_group.
	 *	This is OK, __unhash_process()->list_del_rcu() does not
	 *	clear the ->next pointer, we will find the new leader via
	 *	next_thread().
	 */
1600
	rcu_read_lock();
1601
	for_each_process(g) {
1602 1603
		if (g == tsk->group_leader)
			continue;
1604 1605
		if (g->flags & PF_KTHREAD)
			continue;
1606 1607 1608
		p = g;
		do {
			if (p->mm) {
1609
				if (unlikely(p->mm == mm)) {
1610
					lock_task_sighand(p, &flags);
1611
					nr += zap_process(p);
1612 1613
					unlock_task_sighand(p, &flags);
				}
1614 1615
				break;
			}
1616
		} while_each_thread(g, p);
1617
	}
1618
	rcu_read_unlock();
1619
done:
1620
	atomic_set(&core_state->nr_threads, nr);
1621
	return nr;
L
Linus Torvalds 已提交
1622 1623
}

1624
static int coredump_wait(int exit_code, struct core_state *core_state)
L
Linus Torvalds 已提交
1625
{
1626 1627 1628
	struct task_struct *tsk = current;
	struct mm_struct *mm = tsk->mm;
	struct completion *vfork_done;
O
Oleg Nesterov 已提交
1629
	int core_waiters;
L
Linus Torvalds 已提交
1630

1631
	init_completion(&core_state->startup);
1632 1633
	core_state->dumper.task = tsk;
	core_state->dumper.next = NULL;
1634
	core_waiters = zap_threads(tsk, mm, core_state, exit_code);
O
Oleg Nesterov 已提交
1635 1636
	up_write(&mm->mmap_sem);

1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649
	if (unlikely(core_waiters < 0))
		goto fail;

	/*
	 * Make sure nobody is waiting for us to release the VM,
	 * otherwise we can deadlock when we wait on each other
	 */
	vfork_done = tsk->vfork_done;
	if (vfork_done) {
		tsk->vfork_done = NULL;
		complete(vfork_done);
	}

O
Oleg Nesterov 已提交
1650
	if (core_waiters)
1651
		wait_for_completion(&core_state->startup);
1652 1653
fail:
	return core_waiters;
L
Linus Torvalds 已提交
1654 1655
}

O
Oleg Nesterov 已提交
1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676
static void coredump_finish(struct mm_struct *mm)
{
	struct core_thread *curr, *next;
	struct task_struct *task;

	next = mm->core_state->dumper.next;
	while ((curr = next) != NULL) {
		next = curr->next;
		task = curr->task;
		/*
		 * see exit_mm(), curr->task must not see
		 * ->task == NULL before we read ->next.
		 */
		smp_mb();
		curr->task = NULL;
		wake_up_process(task);
	}

	mm->core_state = NULL;
}

1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725
/*
 * set_dumpable converts traditional three-value dumpable to two flags and
 * stores them into mm->flags.  It modifies lower two bits of mm->flags, but
 * these bits are not changed atomically.  So get_dumpable can observe the
 * intermediate state.  To avoid doing unexpected behavior, get get_dumpable
 * return either old dumpable or new one by paying attention to the order of
 * modifying the bits.
 *
 * dumpable |   mm->flags (binary)
 * old  new | initial interim  final
 * ---------+-----------------------
 *  0    1  |   00      01      01
 *  0    2  |   00      10(*)   11
 *  1    0  |   01      00      00
 *  1    2  |   01      11      11
 *  2    0  |   11      10(*)   00
 *  2    1  |   11      11      01
 *
 * (*) get_dumpable regards interim value of 10 as 11.
 */
void set_dumpable(struct mm_struct *mm, int value)
{
	switch (value) {
	case 0:
		clear_bit(MMF_DUMPABLE, &mm->flags);
		smp_wmb();
		clear_bit(MMF_DUMP_SECURELY, &mm->flags);
		break;
	case 1:
		set_bit(MMF_DUMPABLE, &mm->flags);
		smp_wmb();
		clear_bit(MMF_DUMP_SECURELY, &mm->flags);
		break;
	case 2:
		set_bit(MMF_DUMP_SECURELY, &mm->flags);
		smp_wmb();
		set_bit(MMF_DUMPABLE, &mm->flags);
		break;
	}
}

int get_dumpable(struct mm_struct *mm)
{
	int ret;

	ret = mm->flags & 0x3;
	return (ret >= 2) ? 2 : ret;
}

W
WANG Cong 已提交
1726
void do_coredump(long signr, int exit_code, struct pt_regs *regs)
L
Linus Torvalds 已提交
1727
{
1728
	struct core_state core_state;
L
Linus Torvalds 已提交
1729 1730 1731 1732 1733
	char corename[CORENAME_MAX_SIZE + 1];
	struct mm_struct *mm = current->mm;
	struct linux_binfmt * binfmt;
	struct inode * inode;
	struct file * file;
D
David Howells 已提交
1734 1735
	const struct cred *old_cred;
	struct cred *cred;
L
Linus Torvalds 已提交
1736
	int retval = 0;
A
Alan Cox 已提交
1737
	int flag = 0;
1738
	int ispipe = 0;
1739
	unsigned long core_limit = current->signal->rlim[RLIMIT_CORE].rlim_cur;
1740 1741 1742
	char **helper_argv = NULL;
	int helper_argc = 0;
	char *delimit;
L
Linus Torvalds 已提交
1743

S
Steve Grubb 已提交
1744 1745
	audit_core_dumps(signr);

L
Linus Torvalds 已提交
1746 1747 1748
	binfmt = current->binfmt;
	if (!binfmt || !binfmt->core_dump)
		goto fail;
D
David Howells 已提交
1749 1750 1751 1752 1753 1754 1755

	cred = prepare_creds();
	if (!cred) {
		retval = -ENOMEM;
		goto fail;
	}

L
Linus Torvalds 已提交
1756
	down_write(&mm->mmap_sem);
R
Roland McGrath 已提交
1757 1758 1759
	/*
	 * If another thread got here first, or we are not dumpable, bail out.
	 */
1760
	if (mm->core_state || !get_dumpable(mm)) {
L
Linus Torvalds 已提交
1761
		up_write(&mm->mmap_sem);
D
David Howells 已提交
1762
		put_cred(cred);
L
Linus Torvalds 已提交
1763 1764
		goto fail;
	}
A
Alan Cox 已提交
1765 1766 1767 1768 1769 1770

	/*
	 *	We cannot trust fsuid as being the "true" uid of the
	 *	process nor do we know its entire history. We only know it
	 *	was tainted so we dump it as root in mode 2.
	 */
1771
	if (get_dumpable(mm) == 2) {	/* Setuid core dump mode */
A
Alan Cox 已提交
1772
		flag = O_EXCL;		/* Stop rewrite attacks */
D
David Howells 已提交
1773
		cred->fsuid = 0;	/* Dump root private */
A
Alan Cox 已提交
1774
	}
1775

1776
	retval = coredump_wait(exit_code, &core_state);
D
David Howells 已提交
1777 1778
	if (retval < 0) {
		put_cred(cred);
1779
		goto fail;
D
David Howells 已提交
1780 1781 1782
	}

	old_cred = override_creds(cred);
L
Linus Torvalds 已提交
1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794

	/*
	 * Clear any false indication of pending signals that might
	 * be seen by the filesystem code called to write the core file.
	 */
	clear_thread_flag(TIF_SIGPENDING);

	/*
	 * lock_kernel() because format_corename() is controlled by sysctl, which
	 * uses lock_kernel()
	 */
 	lock_kernel();
1795
	ispipe = format_corename(corename, signr);
L
Linus Torvalds 已提交
1796
	unlock_kernel();
1797 1798 1799 1800 1801 1802 1803 1804
	/*
	 * Don't bother to check the RLIMIT_CORE value if core_pattern points
	 * to a pipe.  Since we're not writing directly to the filesystem
	 * RLIMIT_CORE doesn't really apply, as no actual core file will be
	 * created unless the pipe reader choses to write out the core file
	 * at which point file size limits and permissions will be imposed
	 * as it does with any other process
	 */
1805
	if ((!ispipe) && (core_limit < binfmt->min_coredump))
1806 1807
		goto fail_unlock;

1808
 	if (ispipe) {
1809
		helper_argv = argv_split(GFP_KERNEL, corename+1, &helper_argc);
1810 1811 1812 1813 1814
		if (!helper_argv) {
			printk(KERN_WARNING "%s failed to allocate memory\n",
			       __func__);
			goto fail_unlock;
		}
1815 1816 1817 1818
		/* Terminate the string before the first option */
		delimit = strchr(corename, ' ');
		if (delimit)
			*delimit = '\0';
1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831
		delimit = strrchr(helper_argv[0], '/');
		if (delimit)
			delimit++;
		else
			delimit = helper_argv[0];
		if (!strcmp(delimit, current->comm)) {
			printk(KERN_NOTICE "Recursive core dump detected, "
					"aborting\n");
			goto fail_unlock;
		}

		core_limit = RLIM_INFINITY;

1832
		/* SIGPIPE can happen, but it's just never processed */
1833 1834
 		if (call_usermodehelper_pipe(corename+1, helper_argv, NULL,
				&file)) {
1835 1836 1837 1838 1839 1840
 			printk(KERN_INFO "Core dump to %s pipe failed\n",
			       corename);
 			goto fail_unlock;
 		}
 	} else
 		file = filp_open(corename,
1841 1842
				 O_CREAT | 2 | O_NOFOLLOW | O_LARGEFILE | flag,
				 0600);
L
Linus Torvalds 已提交
1843 1844
	if (IS_ERR(file))
		goto fail_unlock;
1845
	inode = file->f_path.dentry->d_inode;
L
Linus Torvalds 已提交
1846 1847
	if (inode->i_nlink > 1)
		goto close_fail;	/* multiple links - don't dump */
1848
	if (!ispipe && d_unhashed(file->f_path.dentry))
L
Linus Torvalds 已提交
1849 1850
		goto close_fail;

1851 1852 1853
	/* AK: actually i see no reason to not allow this for named pipes etc.,
	   but keep the previous behaviour for now. */
	if (!ispipe && !S_ISREG(inode->i_mode))
L
Linus Torvalds 已提交
1854
		goto close_fail;
I
Ingo Molnar 已提交
1855 1856 1857 1858
	/*
	 * Dont allow local users get cute and trick others to coredump
	 * into their pre-created files:
	 */
1859
	if (inode->i_uid != current_fsuid())
I
Ingo Molnar 已提交
1860
		goto close_fail;
L
Linus Torvalds 已提交
1861 1862 1863 1864
	if (!file->f_op)
		goto close_fail;
	if (!file->f_op->write)
		goto close_fail;
1865
	if (!ispipe && do_truncate(file->f_path.dentry, 0, 0, file) != 0)
L
Linus Torvalds 已提交
1866 1867
		goto close_fail;

1868
	retval = binfmt->core_dump(signr, regs, file, core_limit);
L
Linus Torvalds 已提交
1869 1870 1871 1872 1873 1874

	if (retval)
		current->signal->group_exit_code |= 0x80;
close_fail:
	filp_close(file, NULL);
fail_unlock:
1875 1876 1877
	if (helper_argv)
		argv_free(helper_argv);

D
David Howells 已提交
1878 1879
	revert_creds(old_cred);
	put_cred(cred);
O
Oleg Nesterov 已提交
1880
	coredump_finish(mm);
L
Linus Torvalds 已提交
1881
fail:
W
WANG Cong 已提交
1882
	return;
L
Linus Torvalds 已提交
1883
}